Wood research : bulletin of the Wood Research Institute Kyoto
University (1997), 84: 130-173
Departmental Bulletin Paper
Curing Behavior and Bonding Performance
of Wood Adhesives under
High-Pressure Steam*Kenji UMEMURA
(Received Mav 31. 1997)
Kqwordl : curing behavior, bonding performance, steam-injection heating, wood adhesives,
phenol-formaldehyde reSll1, urea-f(xmaldeDvde reSll1
Contents I n trod uction
Part 1 Curing Behavior of Phenol-Formaldehyde Resin under High-Pressure Steam 1.1 Introduction
1.2 Chemical analysis of curing behavior 1.2.1 Materials and methods
1.2.2 Results and discussion
(1) Curing reaction of 2-hydroxybenzyl alcohol
(2) Curing behavior of PI" resin examined by FT-IR, ]'C-NNIR (3) Insoluble matter of steam-injection heated reS1l1
1.3 Dynamic mechanical properties 1.3.1 :Materials and methods 1.3.2 Results and discussion
(1) Curing process of liquid PI' reSlll
(2) Curing behavior of resin pre-heated with hot-platen heating
(3) Dynamic mechanical properties of reS1l1 pre-heated with steam-il~jection
heating 1.3.3 Summary
Part 2 Curing Behavior of U rca-Formaldehyde Resin under High-Pressure Steam 2.1 Introduction
2.2 Chemical analysis of curing behavior
* This review article is the abstract of the doctor thesis In the author (Kvoto University, 1996). ** Laboratory of Structural Function.
2.2.1 Materials and methods 2.2.2 Results and discussion
(1) Curing behavior of UF reSIn examined by FT-IR (2) Insoluble matter of steam-injection heated resin 2.2.3 Summary
2.3 Dynamic mechanical properties 2.3.1 Materials and methods 2.3.2 Results and discussion
(1) Curing process of liquid UF reSIn
(2) Dynamic mechanical properties of resin pre-heated with hot-platen heating (3) Dynamic mechanical properties of resin pre-heated with steam-injection
heating 2.3.3 Summary
Part 3 Bonding Performances of Wood Adhesives under High-Pressure Steam 3.1 Introduction
3.2 Bonding strength of PF reSIn 3.2.1 Materials and methods 3.2.2 Results and discussion
(1) Effects of additives and steam-injection time on bonding strength (2) Changes in glue line temperature
(3) Improvement of bonding strength 3.2.3 Summary
3.3 Bonding strength of UF reSIn 3.3.1 Materials and methods 3.3.2 Results and discussion
(1) Effects of additives and steam-injection time on bonding strength (2) Changes in glue line temperature
(3) Improvement of bonding strength 3.3.3 Summary
Conclusion Acknowledgment References
In the production of wood based materials, the pressing is one of the most important processes for both prod uctivity and quality of products. Instead of using conventional hot-platen heating, recently, several pressing methods have been developed and introduced!). In particular, steam-injection pressing method in which high-pressure steam is injected into
-131-wood materials during pressing, has received much attention all over the world:!). Many researches on wood based materials using steam-injection pressing have been reported:l-'l). In the steam-injection pressing, as the wood materials are exposed to high-pressure steam, the temperature of wood elements rises immediately compared to hot-platen pressing, which accelerates the cure of resin adhesives. As a result, the pressing time becomes shorter than by conven tional hot -pIa ten pressing:!).
Almost all of the adhesives currently used to manufacture wood materials are formaldehyde resin adhesives. It is very important to know the curing characteristics of adhesives during the pressing processes. However, researches on adhesives have been done mainly on the assumption that wood products are manufactured by hot-platen pressing. Though some researchersh-10) have reported the effect of high-pressure steam on curing of adhesive, they did not discuss in detail. Consequently, the steam-injection pressing conditions for the manufacture of wood composite materials have been decided empirically so far. Therefore, the basic information on curing process and bonding performance of adhesives under high-pressure steam is necessary for the development of the steam-injection pressing technology.
The purpose of present study is to clarify the curing behavior and to establish the optimum bonding conditions of wood adhesives, particularly PF and UF resins, under high-pressure steam. In the first place, the curing mechanisms ofPF and UF resins under high-pressure steam were discussed. That is, resins treated by steam-injection heating and hot-platen heating were analyzed and compared with each other using techniques such as fourier transform infrared spectroscopy (FT-IR), solvent extraction and dynamic mechanical analysis (DMA). In the second place, based on the obtained results, the bonding performance and optimum bonding conditions of each resin under high-pressure steam were investigated.
Part 1 Curing Behavior of Phenol-Formaldehyde Resin under High-Pressure Steam
The curing reactions of PI" resol resin greatly depend on the formaldehyde/phenol mole ratio, pH of the resin, and the reaction temperature. As shown in Fig. 1.1, the cure of PF rcsol resin involves the substitution reactions of the methylol reactive group at a reactive site of an aromatic ring yielding a methylene linkage between two aromatic rings (Reaction 1).
Italso can be speculated that a substitution reaction by a methylol reactive group of another methylol hydroxyl group yielded a methylene ether linkage (Reaction 2). However, the primary reaction under alkaline conditions is considered to be the formation of a methylene linkage. From extensive investigation, minor ether-forming reaction was found to occur,
OH OH OH OH
o-CH,OH-- o-CH,--6-CH,oH+ H20 (1)
OH OH OH OH
- . .
o-CH,-O-eH,-o+ H20 (2)
o-CHO+ (4) (5) o II
Fig. 1.1. Curing reactions of resol-type phenol formaldehyde resin.
and methylene ether linkages formed seem to break down with longer periods of cure or higher cure temperatures (140-170°C)II). The three possible products of decomposition of the ether linkage that have been received as represented by Reactions (3) to (5). That is, the ether linkage splits to give a methylene bridge (Reaction 3) or decompose to give a phenolic aldehyde and methylated phenol (Reaction 4) or decompose to form a quinone methide structure (Reaction 5). Reaction (3) however, is considered the main reaction, and Reactions (4) and (5) are considered to be side reactions12) .
The objective of this part is to reveal the curing behavior of PF reSIn under high-pressure steam chemically and physically. The section 1.2 dealt with the curing behavior of PF resol resin under high-pressure steam, using a specially designed reaction ce1l8). First,
using 2-hydroxybenzyl alcohol as a model compound, the differences between the curing reactions under hot-platen heating and under steam-injection heating were observed. Second, the PF resins heated by both two heating methods were analyzed by FT-IR, 13C_
NMR and solvent extraction.
The section 1.3 dealt with the curing process and the effects of the pre-heating of PF resin under both heating conditions measured by DMA. In addition, the dynamic mechanical curing index, indicating the degree of dynamic mechanical cure for the pre-heated PF resins, was proposed, and its behaviors were evaluated.
-1.2 Chemical analysis of curing behavior 1.2.1 Materials and methods
( 1) Materials
The aqueous PF resol resin was prepared by initially reacting 94 g of phenol (P) with 162 g offormaldehyde solution (F), to give an F/Pmolar ratio of2.0. Thirty-two grams (32 g) of50% NaOH solution was also added as a catalyst, to give a NaOH/P ratio of 0.4. The mixtures were heated gradually until 75°C for 60 min and then were heated for 5 h at 76°C. Upon reaching a viscosity of 1.1 poise, the solution was cooled to 20°C for 90 min. The properties of PF resin were shown in Table 1.1. The model compound, 2-hydroxybenzyl alcohol (Aldrich Chern. Co.) was dissolved in a NaOH solution. The solution had a 46% solid content with pH II, as well as PF resol resin.
(2) Heating methods
Steam-injection heating was done by using a specially designed experimental apparatus as shown in Fig 1.2i\). Each adhesive was weighed and contained in a
polytetra-ftuoroethylene film bag. The amounts of resin were about 0.2 g for FT-IR and about 2.5 g
Tahle 1.1. Properties of PF reSIn.
Viscosity (25°C) Specific gravitv (25°C) Molecular weigh (i\!ln)
Gelation time (hoi ling water) Non-volatile component (135°C, I h) II 0.088(P:n) O.ln 705 27.27 (min) 45.8 (%)
Legend: Mil: Numh_er average molecular weight. Note: Molecular weight (Mil) obtained Ii·om polythylene glycol calibration curve of gel permeation chromatography.
Q9 :Ball valves
for other analyses. The procedure followed the suggestion of Subiyanto et alB). That is, the polytetrafluoroethylene film bag was tied loosely with wire immediately before putting them in the reaction cellSQas to prevent the condensed water from penetrating into the bag
and also to prevent the bag from bursting from adhesive reaction or boiling. The temperature inside the reaction cell was recorded at certain time intervals. After a few seconds of warming, condensed water was drained in the steam line by opening the ball-valves D and E. Valve E was then closed and a high-pressure steam was injected into the reaction cell by opening valve A. The target heating temperature was achieved by controlling steam pressure from a steam generator. The heating temperatures (steam pressure) were 160°C (0.63 MPa) with steam-injection times from 10 s to 60 min. After steam injection for the scheduled time, valve A was closed and valve B was opened to reduce steam pressure for a few seconds followed immediately by opening valve C to inject liquid nitrogen (Liq. N2 ) into the reaction cell until the adhesives were frozen. Frozen samples
were taken out and kept in a freezer at - 20°C until analysis was done.
In hot-platen heating, about 0.2 g (for FT-IR) and 2.5 g (for other analyses) of resin contained in separate polytetrafluoroethylene films were placed between the platens (about 2.5 mm opening between upper and lower platens). The heating temperatures were 160°C with heating times from 10 s to 60 min. After the resin was heated for the scheduled time, it was dipped in liquid nitrogen in order to stop the curing reaction. Then, all the samples were kept in a freezer set at -20°C.
(3) FT-IR and solvent extraction
The frozen samples were vacuum-dried and ground to a powder before analysis. Infrared absorbance spectra was obtained on a Nihon Bunko FT/IR-7000 using KBr pellet technique. To observe the reaction quantitatively, intensity ratio method was used for each obtained spectrum.
Solid-state 13C_NMR spectra was obtained at 50.3 MHz with a Varian XL-200 spectrometer equipped with a CP-MAS probe made by Doty Scientific Inc.
A heat treated and pulverized (less than 20 mesh) dry sample weighing 2.5 g was extracted for 24 h with methanol in a Soxhlet extractor. After that, the amount of residue was determined, and the proportion of the insoluble matter was calculated.
1.2.2 Results and discussion
(1) Curing reaction of 2-hydroxyben?J!l alcohol
The curing reaction of 2-hydroxybenzyl alcohol as a model compound was observed. Fig. 1.3 shows the IR spectral changes of this compound under hot-platen heating (A) and steam-injection heating (B). These spectra were measured after heating the samples at 160°C for 1 min (a), 5 min (b), and 20 min (c) respectively.
In hot-platen heating (A), the curing process of this compound was characterized in a previous report as follows13) : The intensity of the band of the compound heated for 30 min
-f(A) Hot-platen C-OH 1100 1000 Wavenumber (cm-1)
fQ) CJ c as .Q
...o I/) .Q c:( (8) Steam-injection C-OH 1100 1000 Wavenumber (cm-1)
Fig. 1.3. Infrared spectra of2-hydroxybenzyl alcohol during hot-platen heating (A) and steam-injection heating (B) at 160°C. Legend: (a) 1 min, (b) 5 min, (c) 20 min.
at 160°C due to the C-OH stretching vibrational mode at about 995 cm-I
, decreases, and the intensity of the band due to the C-O-C stretching vibrational mode at about 1070 cm- I also
decreases. In this study, the band due to the C-OH mode at about 995 cm- I rapidly decreased, and that due to the C-O-C mode at about 1,070 cm- I increased with increasing heating time under hot-platen heating. Besides Reaction (1) as shown in Fig. 1.1, the Reaction (2) could occur at 160°C for 20 min under hot-platen heating.
In steam-injection heating (B), the band at about 995 cm- I decreased, and the band at about 1,070 cm- I increased and then decreased with increasing heating time. As the formation and decomposition of ether structure were observed, Reaction (2)-(5) were considered to occur in steam-injection heating. However, the formation of the peaks arising from aldehyde group and quinone methide structure were not observed in the IR chart. Therefore, besides Reaction (I),Reaction (2) and (3)would occur at 160°C for 20 min under steam-injection heating. That is, in the curing of 2-hydroxybenzyl alcohol under steam-injection heating at 160°C, the ether linkage was thought to decompose rapidly.
(2) Curing behavior of PF resin examined by FT-IR
Based on previous works on IR measurements of PF resin14,15), the bands of primary
interest are about 1,0IOcm-1
for CoO stretch ofmethylol (CH2-OH),
for CoO stretch of ether (C-O-C), and 1,650cm-1
for C=O stretch of carbonyl (C=O). To observe the curing behavior of the resin quantitatively, as well as to compare the heated resins which were obtained by the two heating methods, the ratios of the intensities of the bands at about 1,010, 1,050, and 1,650 cm- I to that of the aromatic C = C stretching
~__ . . . t l
-tc=o (A) Hot-platen 1.2 CH2-0H
S --C-O-C M O 5 10 15 20 60
Heating time (min) (B)Steam-injection 1.2 CD U c as .Q
o0.8 ~as CD > "iii 0.4 Gi a: ~ --ll---ll
tC=O 5 CH2-0H
t- - --e---II---e L...,j-.~ ---AL.-~--'-~--JlL---l~- -C-O-C 60 CD U c as ~ 0.8 III .Q as CD ~ 0.4 as Gi a: 10 15 20
Heating time (min)
Fig. 1.4. Infrared changes of PF resin during hot-platen heating (A) and steam-injection heating (B) at 160°C. Legend:.: 1,650 cm-I/1,61O em-I, . : 1,010 cm- I/1,610 em-I, ... : 1,050 cm- I/l,610 em-I.
vibrational mode at about 1610 cm-I
were plotted against the heating time16). The results are shown in Fig. 1.4. By hot platen heating (A) in Fig. 1.4, the ratio of 1,650 cm-I
gradually increased with increasing heating time, while 1,010 cm-I
rapidly decreased for 3min and then gradually decreased. The ratio of 1,050 cm-I
/l,610 cm-I slightly decreased over the entire heating time. By steam-injection (B), 1,650 cm-I
rapidly increased and soon reached a constant value within 20 min, while 1,010 cm-I
decreased rapidly in the early stage (to3 min), then maintained a constant value. 1,050 cm-I
also rapidly decreased, and finally it almost kept its value of 0.
Fig. 1.5 shows the 13C_NMR spectra of the hot-platen heated resin (A) and that of the steam-injection heated resin (B) at 160°C for 20 min. The structural assignments were derived from the literature 17,18). In hot-platen heated resin (A), about 62 ppm of methylol and about 73 ppm of ether linkage were observed. In steam-injection heated resin (B), these peaks had weaker intensities than that in hot-platen heating, especially, the peak of ether linkage. This tendency agreed with the IR data in Fig. 1.4. Consequently, the curing behavior of PF resin under steam-injection heating at 160°C seemed to be accompanied by a rapid breakdown of the ether linkage. With regard to the breakdown of the ether linkage, there are three possible pathways shown in Fig. 1.1. From the results of FT-IR and 13C_NMR of each heated resin, we could not confirm the products formed in
Fig. 1.5. 50.3 MHz CP/MAS NMR spectra of heated PF resin. Legend: (A) : Hot-platen heating at 160°C for 20 min, (B) : Steam-injection heating at 160°C for 20 min. Note: Arterisks indicate position of spinning sidebands.
Reactions (4) and (5). So, besides Reaction (1), it seemed that the curing reaction of resin under steam-injection heating may occur mainly as shown by Reaction (3).
(3) Insoluble matter of steam-irljection heated resin
Figs. 1.6 show the result of extraction test with methanol on the resins heated by both heating methods. In the case of resin heated by hot-platen, the methanol insoluble matter gradually increased for 2 min and soon reached almost 100%, whereas rapid a increase was observed for 1 min and then, constant to approximately 90% for the resin heated by
steam-100 ;? ~ 80 ~ Q)
:=tIS '60 E Q)
:c::::s 40 '0 CIl .5 _ 1 1 1 ---,...::....,,---~--II
•- - : Hot-platen heating
o1 2 3 4 5 10
Heating time (min)
Fig. 1.6. Insoluble matter of PF resin during hot-platen heating and steam-injection heating at 160°C. Note: Solvent was methanol.
mJection. These suggested that by steam-injection heating, PF resin reacts more quickly to grow methanol-insoluble matter than by hot-platen heating, and soon the curing reaction is inhibited.
The curing behavior of PF resin under high-pressure steam was studied using several analytical techniques. As heating time increased, curing reaction of 2-hydroxybenzyl alcohol under steam-injection heating may be followed by the formation and then breakdown of ether group. Under hot-platen heating, only the formation of ether groups was observed. The reaction ofPF resin under steam-injection heating was accompanied by the rapid disappearance of ether group. Solvent extraction tests revealed that the methanol insoluble matter formed quickly under steam-injection heating. IR spectra of the methanol extract from steam-injection heated resin showed the formation of carbonyl groups which seemed to be derived from by-products. That is, at 160°C, PF resin immediately cured to some degree in the early stage because of the great heat capacity of steam. After that, however, the curing reaction was inhibited by the condensed water of steam and the by-products in heated resin.
1.3 Dynamic mechanical properties
1.3.1 Materials and methods
The aqueous PF resol resin used in this experiment is the same as described in section 1.2.1 (1).
(2) Preparation of specimen and pre-heating methods
A total of strip of glass fiber filter (ADVANTEC GA-200), measuring 70 X8X 0.75 mm (about 0.1 g), was dipped in ajar ofPF resin in a brief time. Upon removal from the jar, the specimen was laid flat on a clean sheet of polytetrafluoroethylene and the excess resin was squeezed out uniformly. About 1 gram of resin was retained by each specimen. The specimens were then pre-heated using hot-platen (in direct contact with upper and lower platens). The heating temperatures was 160°C, with heating times from 1 min to 120 mm. After the specimen was heated for the scheduled time, it was dipped in liquid nitrogen in order to stop the curing reaction. Then, all the specimens were kept in a freezer set at -20°C.
In pre-heating using steam-injection, the specimens were held between two polytetrafluoroethylene, and were inserted into a thimble filter so as to prevent the resin from being washed out. The thimble filter containing the specimen was put into the reaction cell. Then, steam-injection heating was performed according to section 1.2.1 (2). The heating temperature at 160°C (steam pressure of 0.63 MPa) was used, with heating times ranging from 0.5 min to 20 min. The specimens were also kept in the freezer at -20°C.
-(3) D MA measurement
The frozen specimen was taken out of the refrigerator, and exposed to ambient temperature for 15 min. After that, DMA measurement was performed with RHEO VIBRON DDV-25FP instrument, scanning from 20°C to 220°C at a rate of 3°C/min and 10 Hz.
1.3.2 Results and discussion
(1) Curing process
ofliquid PF resin
Fig. 1.7 shows the changes of dynamic mechanical properties during the curing process of liquid PF resin. The value ofE increased gradually in oven temperature up to 114°C (tl)' After that, it increased dramatically until 135°C (t2), and soon reached a constant value. The values of
0had peaks at 129°C and 123°C (1), respectively. Fukuda et al.19) measured the curing process of dehydrated PF resol resin using torsional
braid analysis (TBA). The behaviors of relative storage modulus (Gr) and logarithmic decrement obtained from the measurement indicated similar behaviors of E' and tan 0as shown in Fig. 1. 7. According to Fukuda et al.19), the point at which Gr began to increase
rapidly (corresponding totl in Fig 1.7) was regarded as the gelation point of the resin. The stop point of the rapid increase in Gr (corresponding to t2) was regarded as the finishing
point of resin curing reaction. The peak oflogarithmic decrement (corresponding to1)was regarded as the phase transition point during curing process.
In this study, as the PF resin solution contained about 54% of water (resin solid content about 46%), the mechanical curing process is considered to be affected by both the resin and water. The curing process of dried PF resin where water was excluded is shown in Fig. 1.8. The value of
Edecreased slightly from the beginning of the measurement till about 55°C
101 ° E' 10 0 E" + -4 tan0 ~!IIIl!!!l(j!!!l!!l!1 10° °i oj 'is
~11-0 a. c ~ 10-1
~~0.1 ~ ~
,"00'"._~ [I, : 10-2
~:i0.01 ' I : ~ i:! T t1 i: :j' :yt2 0.001 10-3 0 50 100 150 200 250 Oven temperature rC)
Fig. 1.7. Dynamic viscoelastic properties of PF resin observed at 3°C/min. Note: 11 :Gelation point, 12: Finishing point of resin curing reaction,
E'during measurement was probably due to the high
E'of glass fiber filter specimen containing dried PF resin, even at room temperature. The
E"and tan 0 had broad peaks from 55°C to 138°C. Itcan be presumed that the marked curing reaction ofPF resin itself takes place from 55°Cto 138°C. Based on the above, the curing process ofliquid PF resin in Fig 1.7 was analyzed as follows.
From 20°C up to 114°C (t1), the curing reaction took place with the formation of linear and short chains. However, as the PF resin has low molecular weight, and contained considerable amount of water at this stage, low values of
E"were recorded. From
10-1 o E' o E" ~ tan 0 <nzan«([fCnIIUIIj(IIH«ItItU(BI((((H((UIIH B,III(I( I +--~
~::::::~~ 10 1 0.1 10-2~~---'-_-'---'-_~--'--~_..L..-~--J 0.01 o 50 100 150 200 250
Fig. 1.8. Dynamic viscoelasticity of dried PF resin during scanning at 3°C/min and 10 Hz. Pre-heating time • Untreated ... 1 min Ii. 2min o 3min o 4min • 20 min
ttPre-heating time Do ~
•Untreated ~ ... 1min 10-1 .. ' " .."I"
""'iIi. 2min 0 3 min 0 4min 1 t1 II 20 min i~ 10-2 0 50 100 150 200 250 t.O I: 0.1 S 0.01 o 50 100 150 200 250
Oven temperature eC)
Fig. 1.9. Storage modulus (E') as a function of oven temperature at an increasing rate of 3°C/min, with a parameter of pre-heating time at 160°C of resin. Note: tl: See Fig. 1. 7.
Oven temperature eC)
Fig. 1.10. Tan 0 as a function of oven tempe-rature at an increasing rate of 3°C/ min, with a parameter of pre-heating time at 160°C of resin.
-141-114°C (td to 135°C (t2),
E"increased drastically. During this period, linear and branched chains continued to grow, and some of the branched chains could at the same time react locally together. In addition, marked evaporation of water took place. Katovlc et al.20) reported evaporation of water above 100°C in PF resin. Therefore, the gelation of the
reSIn was thought to take place at this stage. Above 135°C, the value ofE' was almost constant. PF resin was immobilized by further curing reaction. Thus, the structure changed to glassy state which retarded further reaction.
Fig. 1.9 shows the
E'behaviors of PF resin samples pre-heated at 160°C using hot-platen for various duration. The values of
E'at 20°C increased with increasing pre-heating time. That is, as the curing of PF resin and evaporation of water took place during pre-heating, the rigidity of pre-heated resin increased gradually. Fig. 1.10 shows the tan 0 behaviors of PF resin samples pre-heated at 160°C using hot-platen for various duration. For both the 1- and 2-min samples, tan 0 had sharp peak at about 120°C. As the pre-heating time became longer, tan 0peak broadened at lower temperatures and became low. For 20 min samples, since no peak value of tan0 was recorded, and the value of
E'was almost constant during the DMA measurement, pre-heated resin was considered to have cured fully.
(2) Curing behavior of resin pre-heated with hot-platen heating
Generally, dynamic mechanical properties of wood adhesives are considered to have a significant influence on bonding performance. However, as wood adhesives are dissolved in solvents, the mechanical curing processes are affected by both resins and solvents. Therefore, calculation method for relative degree of mechanical cure involving the effect of solvents was investigated.
Fig. 1.11 shows the influence of pre-heating treatment on a typical
E'behavior during DMA measurement at a constant heating rate. In the case of the untreated liquid resin (A), the
E'value at the beginning of measurement (20°C) was the lowest. It increased gradually in lower oven temperature range. After that,
E'dramatically increased and soon reached a constant value. In the case of the incompletely cured resin (B), it somewhat softened as the oven temperature increased, indicating a decrease of the E'. Following this, further curing reaction occurred, whereby
E'eventually rose to a plateau. In the case of the fully cured resin (C), the value of
E'was almost constant during the entire measurement because of the good thermostability of the resin. That is, when DMA measurement of pre-heated resin is performed from ambient temperature to elevated temperature at a constant heating rate, pre-heated resin would be finally cured completely at elevated temperature. The relative degree of mechanical cure achieved in pre-heated resin was measured based on the variation in
E',as discussed below.
The lowest In
E'value of the untreated liquid resin at the beginning of DMA measurement (20°C) is represented by In E'L. The highest In
E'value of the untreated
. (C) Full cured sample
(E'H -E'L) 0 (B) Incompletely cured resin sample
(A) Liquid resin sample
...-- _- _-.----..---_ ---
._--Oven temperature eC)
Fig. 1.11. Influence of pre-heating treatments on storage modulus curves.
resin at a elevated temperature is represented by InE'H. The difference between InE'Hand In £ L is expressed as:
(In £H-ln £rJo=ln (£H/E'rJo (1)
The expression in equation (1) indicates the ratio of the highest In £ and the lowest In £ during DMA measurement. Similarly, the lowest In E' value of the pre-heated resin at a low temperature is represented by In £L. The highest In £ value of the pre-heated resin at elevated temperature is represented by In £ H. The difference between In £ H and InE'L is expressed as:
(In £H-In£dt=ln (£H/E'r.)t (2)
The degree of mechanical cure achieved following equation (3), which is derived Dynamic mechanical curing index
in the pre-heated resin is calculated from the from equations (1) and (2):
In (E'H/£do-In (£H/E'r,)t
In (E'H/£do X 100 (3)
£ H: Maximum value of £ at high temperature £ L: Minimum value of £ at low temperature Suffix 0: Untreated
Suffix t: Pre-heating time
Fig. 1.12 shows the dynamic mechanical curing behavior of PF resin based on eq uation (3). The data of insoluble matter extracted with methanol, which is the conventional method, is shown in the same figure. In the case of the dynamic mechanical curing index, the value increased rapidly. The value reached 100 after 5 min, suggesting that resin heated for more than 5 min at 160°C was fully cured. In the case of the methanol
...---- 100 Cii
1~ U 75 75 C ~ tU>C
U1:J Dynamic mechanical
CDC tU e- 50 curing index 50
eCl .2.5 ----0- :160·cheating CD e'"
i:J C U 25 Insoluble matter 25 '0 >- UI C - - . - - : 160'Cheating .5 0 0 0 5 10 15 20
Heating time (min)
Fig. 1.12. Mechanical curing indexes and insoluble matter as functions of heating times. Note: The datas of imsoluble matter were quated from Figs. 1.6.
extraction, the insoluble matter at 160°C increased rapidly for 2 min and soon reached almost 100%. These results are slightly different from that of the dynamic mechanical curing index. The main reason could be due to the change in the size of insoluble matter which was related to the resin molecular size during the curing process. On the other hand, the dynamic mechanical curing index was related to the change of the£ of the resin sample containing water, during the curing process. In addition, this difference also seemed to be influenced by the extraction power of methanol. However, as described previously, dynamic mechanical properties of wood adhesives are considered to have a significant influence on bonding performance. Therefore, dynamic mechanical curing index appeared to be one of the useful methods to evaluate the degree of cure of resin.
(3) Dynamic mechanical properties of resin pre-heated with steam-injection heating
Fig. 1.13 shows the changes in the dynamic viscoelasticities for PF resin treated by steam-injection heating for various times at 160°C during temperature scans up to 220°C. The behavior of all the heated samples under steam-injection heating differed from those under hot-platen heating. As a whole,
E'of steam-injection heated resin gave low value, comparing with that of hot-platen heated resin. The reason seems that the resin contained specimen took place swelling and dilution during steam-injection heating. In the sample pre-heated for 1 min, a slight increase in
£was observed at a lower temperature range (20-80°C), with steady increases in£' and tan 8. Above80°C, further polymerization and crosslinking presumably occurred, whereby£ continued to rise gradually. The value of£' and tan 8 decreased steadily when temperature rose up to 150°C and then somewhat increased. In other samples pre-heated for 5, 10, and 20 min,
£indicated slight decreases in thermal softening at lower temperatures. A slight increase in
£began at about 100°C while sudden dips ofE" and tan 8 were observed at the same temperature. The above phenomena might indicate considerable evaporation of the free water contained in the samples. That is, during the curing of PF resin under steam-injection heating, condensed
101 101 0 10 0 10 1 min E' 10 min E' 0 E" 0 E"
•tan" 100 _"iIQ!~rnii!i!lii"iiIIil 1 ... 100 ~Ult«"'i!!IiI!!ii!i!ii!l!!!!Iiil!!!!1 Ii l'lI c.o a. c.o a. ~ c ~ c S S ~ ~ 10.1
----.0.1 10·1 + - - 0.1
+I!\~10.2 0.01 10·2 0.01 0 50 100 150 200 250 0 50 100 150 200 250
Oven temperature eC) Oven temperature eC)
101 101 10 0 10 5mln 0 E' 20 min E' 0 E" 0 E"
•tan" 100 100 ~(Iiijili!!l!lIi!!!!"ii!!!!i!)
Iia. c.o 'iia. c.o
~ c ~ c ~ S ~ S 10.1 0.1 10.1 0.1 tI *'*t' ----+ ----+
~rilt: ~lIilililiiliillii!liI!i"1 ~ 0.01 10-2 0.01 10-2 0 50 100 150 200 250 0 50 100 150 200 250
Oven temperature ('C) Oven temperature ('C)
Fig. 1.13. Dynamic viscoelasticities of PF resin exposed to 160°C for various heating times during steam-injections.
water from steam and by-products from condensation reactions in PF resin might have been formed, and was confined in the sample due to the high-pressure steam. Thoman and Pearson?) suggested that polymerization of phenolic resin yielded water as a by-product and the additional water from condensation of steam shifts the chemical equilibrium, reducing the resin curing reaction. Gyeviet al.21) also reported that the by-products accumulated in
a closed system could hinder the reaction of PF resin, and led to an equilibrium reaction. Consequently, the water arising from steam and a condensation reaction seemed to hamper the condensation reaction physically. Therefore, similar behavior might have been indicated by the dynamic viscoelasticity of steam-injected samples.
In all cases, it was revealed that the samples pre-heated under steam-Injection has cured to some degree. The degree of mechanical curing was determined by calculating the dynamic mechanical curing index. The results are shown in Fig. 1.14. For the sample heated by hot-platen quoted from Fig. 1.12, the dynamic mechanical curing index increased gradually for 5 min and soon reached almost 100 i.e., fully cured, whereas in the sample heated by steam-injection, it rapidly increased for 1 min and then leveled off at
•--.- : Steam-injection heating - - : Hot-platen heating 20 15 10 5 100 lU (,) 80 C as>< J:CD (,)'t:J 60 CDC
E';:40 as:::J c(,) >-C 20 0 0
Heating time (min)
Fig. 1.14. Dynamic mechanical curing indexes for steam-injection heated and hot-platen heated PF resins as functions of heating time at 160°C.
approximately 85. These behaviors agreed with the results of methanol extraction tests shown in Fig. 1.6. It was also recognized ~hatPF resin cured to some degree immediately in the early stage, and soon maintains a constant value under steam-injection heating. As for the relationship between moisture content and the degree of cure in PF resin, Chow and Mukai6) pointed out that the degree of cure of the resin is directly related to the moisture
content of the resin. Therefore, it can be concluded that steam-injection heating accelerates the curing reaction of PF resin to some degree in the early stage because of the great heat capacity of steam, but tends to hinder the curing reaction thereafter, due to the presence of water. It was for this reason that some free water was retained in PF resin heated under high-pressure steam, hence reducing the condensation reaction so as to decrease the collisions of functional-groups. Furthermore, by-products remained in the heated resin might also hinder the curing reaction.
The curing process of PF reSIn pre-heated with hot-platen and steam-Injection was studied by DMA. The results and conclusion are summarized as follows: Itappeared that the untreated liquid PF resin has a distinct gel point indicated by sudden changes ofE, E",
and tan 0, whereas, dried PF resin exhibited different behaviors. As the pre-heating time was prolonged in hot-platen, there was a general increase in the initial (20°C)E,followed by a drop of tan
0peak. I t was also observed that the resin pre-heated for 5min at 160°C cured completely with the disappearance of the tan 0 peak. In addition, dynamic mechanical curing index which is a method of determining the degree of cure of resin in dynamic mechanical aspects was proposed. The method was considered to be one of the useful methods to evaluate the degree of resin cure. On the other hand, the dynamic viscoelasticity of steam-injected samples indicated similar behavior regardless of a heating time. Even in the resin pre-heated with steam-injection for 20 min at 160°C, a slight increase ofE began at 100°C, together with sudden dips ofE'and tan
o.This could be due to the presence of considerable amount of free water in the samples, Based on the behavior
of dynamic mechanical curing index of steam-injected resin, it was recognized that the PF resin cured immediately to some degree in the early stage, and soon leveled off at a constant value.
PART 2 Curing Behavior of Urea-Formaldehyde Resin under High-Pressure Steam
As wood adhesives, UF resins have been commonly used for wood products using hot-platen pressing. However, in general, UF resins are less stable in humid conditions and at higher temperatures. Many researchers have studied the curings and degradations of UF resins using several analytical methods22-27).
As shown in Fig. 2.1, when the acidic hardening agents and the heat are added, UF resin rapidly becomes gels, producing a highly cross-linked insoluble polymer28). The
formation of this polymer network is a consequence of the polyfunctional properties of urea whose four amide hydrogen atoms are capable of reacting with the hydroxyl groups of methylene glycol which is the hydrate of formaldehyde existing in an aqueous solution (Reaction 1). In the curing· process, possible pathways of the reactions involve the formation of methylene bridges between the methylol and amide groups (Reaction 2). The formation of ether linkages between methylol groups also occurs (Reaction 3). In addition to the above reactions, the formation of uron expressed in Reaction 5 is confirmed. However, the curing reactions of UF resin are greatly affected by the molar ratio between
- N-e-NH + HOCH2OH .. .. - N-C-N-CH20H + H2O (l)
I I I I
0 0 0 0
II II II II
-N-e-NCH20H + HN-e-N-.. • -N·e-NCH2N-e-N-- + H2O (2)
I I I I I I I I
0 0 0
II II II
--N-e-NCH2OCH2·N-C-N - + H2O (3)
-I I I I
-N-e-NCH2(OCH2)nOH + HOCH20H .. ~ -N-C-NCH2(OCH2)n+l0H + H2O (4)
I I 1 I o II -N-C-N-I I CH2 CH2 I I OH OH
Fig. 2.1. Urea-formaldehyde polymelization and cure reaction.
-147-urea and formaldehyde, kind of catalyst, pH of the resin solution, and temperature of the curing reaction. Consequently, the structure is still ambiguous.
The objective of this part is to reveal the curing behavior of UF resin under high-pressure steam both chemically and physically. The section 2.2 deals with the curing behavior of UF resin under high-pressure steam. That is, the UF resins heated by both heating methods were analyzed by FT-IR and solvent extraction.
The section 2.3 deals with the curing process and the effects of the pre-heating of UF resin under two heating methods measured by DMA. In addition, the dynamic mechanical curing behaviors of UF resin under each heating condition were investigated.
2.2 Chemical analysis of curing behavior
2.2.1 Materials and methods
The UF resin was prepared by initially reacting 120 g of urea (U) with 324 g of formaldehyde solution (F), to give a FlU ratio of2.0. The mixture was adjusted to a pH of 7.8-8.2 to have an alkaline solution, and then was heated while continuously stirring for 3.7 h at 85°C. In the course of this reaction, urea was added to bring the resin to a final FlU
ratio of 1.7. The solution was then cooled and neutralized to pH 7.0. The resin properties obtained are shown in Table 2.1. The solution as well as the obtained UF resin had about 50 % solid content. Twenty percent (20%) ammonium chloride aqueous solution was added to the UF resin as a catalyst, at I% based on solid resin weight prior to testing.
(2) Heating methods
Steam-injection heating and hot-platen heating were performed according to the section 1.2.1. The heating temperatures were 160°C (steam pressure of 0.63 MPa) and 120°C (steam pressure of 0.20 MPa) with heating times from 10 s to 20 min. After preparation, all the samples were kept in a freezer at -20°C. The frozen samples were vacuum-dried and ground to a powder before analyses.
(3) FT-IR and solvent extraction
In FT-IR measurement, the infrared absorbance spectra were obtained on a Nihon
Table 2.1. Properties of UF resin. pH Viscosity (25°C) Specific gravity (25°C) Molecular weight (Mn ) Gelation time (25°C) Non-volatile component (105°C, 3 h) 7.0 0.04 (Pa's) 1.180 595 50 (min) 50 (%)
Legend: Mn : Number average molecular geight. Note:
(Mn ) obtained from poliethylene glycol calibration curve of gel permeation chromatography.
Bunko FT-IR 7000 using the KBr pellet technique.
In solvent extraction measurement, heat treated and pulverized (less than 20mesh) a dry sample weighing 2.0g was extracted for 3h with warm water at 60°C. Then, the amount of the residue was determined, and the ratio of insoluble matter was calculated.
2.2.2 Results and discussion
(1) Curing behavior of UF resin examined by FT-IR
Fig. 2.2 shows the infrared spectral changes of UF resin under hot-platen heating (A) and steam-injection heating (B) at 160°C. These measurements were made after heating the samples at 160°C for 1min (a), 5min (b) and 20 min (c), respectively. In hot-platen heating (A), besides a decrease of the absorption intensity of the band at 1,020-1,025cm-I
, slight increases of the absorption intensity of the bands at 1,500-1,520cm-I
, 1,290-1,300 cm-I
and at about 810 cm-I
were observed with increasing heating time. However, dramatic changes did not occur because UF resin rapidly formed highly crosslinked polymer at higher temperature of 160°C.
On the other hand, with steam-injection heating (B), marked changes occurred in the spectra during heating. That is, absorption of the band at about 1,540-1,560cm- I was
missing, and only a strong absorption of the band at about 1,500-1,520cm- I appeared with increasing heating time. Strong absorption of the band at about 1,290-1,300cm- I also appeared. The absorption intensity of the band at about 1,020-1,025cm-I
decreased, and the absorption of the band at 810cm- I was observed clearly.
These absorption bands was assigned as shown below29,30); 1,540-1,560cm-I for
(A) Hot-platen (B) Steam-injection
1CI) u C «I .Q
...o VI .Q <C 1500 1000
1CI) u C «I .Q
...o VI .Q <C 1500 1000 Wavenumber (cm-1) Wavenumber (cm-1)
Fig. 2.2. Infrared spectra of UF reSIn during hot-platen heating (A) and steam-injection heating (B) at l60°C. Legend: (a) 1 min, (b) 5 min, (c) 20 min.
-149-16 20 0-0, 0-0-0,
°o-===-=-o0 - 0 , 0 - 0 - 0 - -~-.--.
o4 8 12
Heating time (min)
o 2 4 6 8 10 20
Heating time (min)
secondary, linear amide; 1,500-1,520 cm-I
for linear or cyclic tertiary amide; 1,290-1,300 cm-I
forOHand/or tertiary cyclic amide; 1,020-1,025 cm- I forCH20H ;810 cm- IforCOC linkages of a uronic ring.
Some researchers29,31,32) tried to observe changes in functional groups during the curing
of UF resin by using obtained infrared spectra data. That is, some of the infrared spectra were quantified on the same relative basis by normalizing the measured absorbances using the carbonyl (at about 1,650 cm-I
) peak as the criterion in each spectrum. To examine the curing behavior of the resin quantitatively, as well as to compare the heated resins treated by the two heating methods with each other, the relative normalized absorbances described above were calculated for each spectrum.
Fig. 2.3 shows the plot of changes in these absorbance against heating times on hot-platen heated (A) and steam-injection heated (B) UF resin at 160°C. In a hot-platen heating (A), the absorption intensity of the ban'd at 1,540-1,560cm-1
decreased slightly, whereas the absorption intensity of the band at 1,500-1,520 cm-I
increased immediately (A) Hot-platen CD o c
iia; IX (8) Steam-injection (8) Steam-Injection
Heating time (min)
IR spectral changes of UF resin during hot-platen heating (A) and steam-injection heating (B) at 120°C. Legend: .... : 1,540-1,560 cm- '/l,650 cm-I, . : 1,500-1,520cm- '/1,650 cm- I, D: 1,290-1,300cm- '/l,650cm- ', 0: 1,020-1,025cm- '/l,650cm- ', . : 810 em -1/1,650 cm -I. Fig. 2.4. CD o C tV .0
..o III .0 tV CD >
. - - . - - -. - 1 1 - .
---_ _ A-.6. .6.-11-.6. ---"-_L..-~-'-~---'-~...L-.~--'---tlt----' 2 4 6 8 10 20
Heating time (min)
IR spectral changes of UF resin during hot-platen heating (A) and steam-injection heating (B) at 160°C. Legend: .... : 1,540-1,560cm- '/l,650 cm -I, . : 1,500-1,520 cm- '/l,650 cm-\ D: 1,290-1,300 cm- '/l,650 cm-I, 0: 1,020-1,025cm- '/1,650cm- ', . : 810 cm -1/1,650 em -I. Fig. 2.3. CD o C tV .0
.gR>-o __0 - 0 - 0 - 1 1 - 0 ~ O.
e=II==eCD ' . IX
after heating and level off soon. The absorption intensity of the band at 1,290-1,300cm-I increased and reached a -maximum in 30s, followed by a considerable decrease. The absorption intensity of the band at 1,020-1,025cm- I decreased soon after treatment and then had a slight increase afterwards. Furthermore, the band at 810 cm-I
began to appear immediately after the heating and level off after 2 min of the treatment. In a steam-injection heating (B), the absorption intensity of the band at 1,540-1,560cm-I
decreased slowly, whereas the absorption intensity of the band at 1,500-1,520cm-I
increased immediately after heating, and then decreased slightly for 2 min, then increased again gradually. The absorption intensity of the band at 1,290-1,300cm-I increased continuously, whereas the absorption intensity of the band at 1,020-1,025cm-I decreased gradually over the entire heating time. In addition, the absorption intensity of the band at 810 cm- I appeared and increased up to30s, slightly decreased at 2 min, and then somewhat
increased. As observed, marked chemical changes seemed to occur within about 2 min in steam-injection heated UF resin.
Fig. 2.4 also shows the plots of changes in some of these absorbance against heating times of hot-platen heated (A) and steam-injection heated (B) UF resin at120°C. Ina hot-platen heating (A), the changes of each absorbance were more slow against heating times than those done at 160°C. Ina steam-injection heating (B), the absorption intensity of the band at 1,540-1,560cm-I
decreased slowly, whereas the absorption intensity of the band at
1,500-1,520cm- I increased significantly at the first stage of heating, decreased largely from 5 min to 10 min heating, and then increased gradually. On the other hand, the absorption intensity of the band at 1,290-1,300cm-I
increased continuously and slightly for 7 min, decreased slightly, and then somewhat increased. The absorption intensity of the band at
decreased slightly over the entire heating time. Furthermore, the absorption intensity of the band at 810 cm- I showed a similar behavior to that of the band at 1,290-1,300cm-I
• Inthis case, marked chemical changes seemed to occur at about 10 min.
Judging from the above, during hot-platen heating of UF resin, tertiary amide grew rapidly accompanied by the consumption of a hydroxyl group, and a uronic ring also seemed to form. These changes were considered to be due to processes such as Reactions (2), (3), (5) of Fig. 2.1 in the forward direction. Similarly, in the steam-injection heating of UF resin, it seemed that the curing reactions occurred rapidly after heating and the UF resin cured to some degree. On the other hand, based on the temporary decrease of tertiary amide (at 1,500-1,520 cm-I
) and the continuous decrease of secondary amide (at
) , the UF resin appeared to be partly decomposed such as a hydrolysis and
so on. Generally, during hydrolysis, Reactions (1) and (4) will occur in the reverse direction. They will be accompanied by the reverse of (2) and (3), and succeeded by the loss of methylol as in (1) and (4)28). However, it was thought that cyclic ether, such as a
-151-uronic ring formed by a curing reaction, was relatively stable even under steam-injection heating.
(2) Insoluble matter of steam-irljection heated resin
Figs. 2.5 and 2.6 show the results of the extraction tests with warm water at 60°C on the resins heated by each heating method at 120°C and 160°C, respectively. In the case of resin heated by hot-platen heating at 120°C in Fig. 2.5, the insoluble matter increased gradually up to 10 min and reached almost 100%, whereas for the resin heated by steam-injection, a rapid increase of the insoluble matter was observed for 1 min and then kept constant reaching approximately 68%. These observations suggested that in steam-injection heating at 120°C, the UF resin reacts more quickly to grow warm water insoluble matter than that in hot-platen heating because of supply of heat energy from steam. However, at a later time, the curing reaction seemed to be hindered by water contained in the resin. In this case, the reactions of UF resin were apparently considered to come to an equilibrium. This behavior was similar to that of PF resin under steam-injection heating at 160°C as described in Part 133).
In Fig. 2.6, irrespective of heating methods, the insoluble matter rapidly increased for a few seconds. After that, in hot-platen heating, the value approached 100%. In steam-injection heating, it decreased gradually with increasing heating time. Itcan be presumed that under steam-injection heating at 160°C the UF resin rapidly cured to some degree in the early stage. However, in general, the UF resin under wet conditions at higher temperatures is less stable. Therefore, hydrolysis would occur which was caused by water included in the resin uncured yet, and by the condensed water from steam. Thus, water soluble matter of low-molecular weight seemed to increase as heating time increased.
20 16 12 Hot-platen heating 8
•Steam-injection heating 4 100 Hot-platen heating a 100 a
;e80 80 ~
•~ 60 E
:cSteam-injection heating Q) ~
:c0 ~ en 0 40
.=20 10 15 20 0 5
Heating time (min)
Fig. 2.5. Insoluble matter ofUF resin during hot-platen heating and steam-injection heating at 120°C. Note: Solvent was warm-water (60°C).
Heating time (min)
Fig. 2.6. Insoluble matter ofUF resin during hot-platen heating and steam-injection heating at 160°C. Note: Solvent was warm-water (60°C).
The curing behavior of UF resin under high-pressure steam was studied using FT-IR and solvent extraction. The results were as follows:
1) In the reaction ofUF resin under steam-injection heating, marked chemical changes took place, which were not observed in the hot-platen heating. Furthermore, the cyclic ether formed by the curing reaction was considered to be relatively stable even under steam-injection heating.
2) Compared with hot-platen heating, UF resin reacted more quickly and warm water insoluble matter increased in steam-injection heating. Afterwards, however, it was shown that the curing reaction of UF resin appeared to come to an equilibrium state at 120°C, and the hydrolysis tended to occur at 160°C.
2.3 Dynamic mechanical properties
2.3.1 Materials and methods
The UF resin used in this experiment is the same as described in section 2.2.1 of this part. Twenty percent (20%) ammonium chloride aqueous solution was added to the resin as a catalyst at 1% based on solid resin weight before specimen preparation.
(2) Preparation of specimens and pre-heating methods
The specimens were prepared by soaking strips (70 X 8 X 0.75 mm) of glass fiber filter in the resin solution before the pre-heating treatment. Dried specimens were prepared by vacuum-drying using a desiccator. The detailed specimen preparation procedure and pre-heating methods were described 1.3.1 (2).
In pre-heating method using hot-platen, the heating temperatures were 120°C and 160°C with heating times from 0.5 to 10 min. In pre-heating method using steam-injection, the heating temperatures were 160°C (steam pressure of 0.63 MPa) and 120°C (steam pressure of 0.20 MPa) with heating times from 0.5 to 20 min.
(3) DMA measurement
DMA was performed with a RHEO VIBRON DDV-25FP instrument, scanning from 20°C to 220°C at a rate of 3°C/min and 10 Hz.
2.3.2 Results and discussion
(1) Curing process of liquid UF resin
Fig. 2.7 shows the changes of dynamic mechanical properties during the curing process of liquid UF resin using ammonium chloride as a catalyst. A double distinct sigmoidal shape having the changing points of To and T2 was observed in the E. Generally,
thermosetting materials, particularly wood adhesives, undergo two significant changes such as gel transition and glass transition during the curing process. However, as wood adhesives are dissolved in solvents, the mechanical curing processes are considered to be
-100 tV Q. ~ 10-1 ~ 10-2 10-3 0 50 100 150 200 250 10 0.1 0.01 Oven temperature ("C)
Fig. 2.7. Dynamic viscoelastic properties of UF resin during the curing process observed at 3°C/min intervals and at 10 Hz.
affected by both the resins and the solvents. In particular, water release occurs during the curing period in aqueous condensation resin, which may cause a phase change in the resin. Toussaint et al.34) measured the curing process of epoxy resin with isothermal DMA
measurements, and also obtained the same behavior of E and E' as shown in Fig. 2.7. They fully discussed this curing process in its morphological aspect. Stuligross and Koutsky 24)reported on the morphological structure of uncured and cured UF resins. They
proposed that the curing of UF resin proceeds via crystallization and agglomeration. Based on the above references, the curing process of liquid UF resin in Fig. 2.7 was analyzed. From the beginning of measurement up to To (about 73°C), as the UF resin was in a liquid state, low values of
Eand E" were observed. In this stage, the resin molecules moved relatively freely in water. The slight increases of the values ofE andE'were mainly considered to be caused by the evaporation of some of the water. From To to T1 (about
80°C), the values ofE and E' increased drastically. During this period, a marked curing reaction of the UF resin appeared to take place, and the movement of the resin molecules was restricted remarkably. Some researchers22,35-37) observed that the curing reaction of
UF resin with ammonium chloride usually began to take place at below 100°C. From T1to
T2 (about 116°C), very little or no variation of the E and E' values were observed. This
result may be attributed to the formation of aggregates or a branched or linear extending chain occurring from some of the microgels of the resin. If this union suddenly increases the molecular weights, its influence on the mean molecular weight between crosslinks is very smaIl34). Consequently, the influence on the E' which depends on the said formation was
considered to be reduced. Inthis stage, however, as some water still appeared to be present in the specimen, it was also necessary to take into account its effects on theE values to some degree. From T2 to T3 (about 188°C), violent evaporation of water appeared to take place
at T2 • Chowet al. 31
) reported that the evaporation of water above lOO°C occurred for UF
reSIn. Furthermore, the numbers, sizes, and shapes of the aggregates, as well as the dimensions of the other growing macromolecules and the degrees of conversions, were such that they reacted with each other. As a result, the structure of the UF resin was immobilized and change to a glassy state. After completion of curing, the resin began to decompose at about T3 as indicated by the decreasing of the E'.
(2) Dynamic mechanical properties of resin pre-heated with hot-platen heating
Fig. 2.8 shows the E' behaviors of UF resin samples after pre-heating at 120°C for various times with hot-platen heating. For these samples, the values ofE' at 20°C increased with increasing pre-heating time. For the 30-s pre-heated sample, as the oven temperature increased, the value of
Itcontinued to increase up to about 150°C. After that, the
Iteventually attained to a "plateau" value. It appeared that the 30-s pre-heated resin did not cure so much, and considerable water was still contained in the sample. The I-min pre-heated sample softened somewhat as the oven temperature increased, indicating a decrease in the E' at about 50°C. Following this, further polymerization and crosslinking pre-sumably occurred, whereby the E' eventually rose to a plateau. For the 5-min pre-heated sample, even though the oven temperature increased, the sample did not softened so much. After that, theE'eventually reached a plateau. As the resin pre-heated for 10 min had been fully cured, the E' of the sample almost maintained a constant value or decreased slightly because of some decomposition of the resin during the measurement.
~ ~ • Untreated 10-1
•~. 00 0.51 6. 5
•10 10-2 0 50 100 150 200 250 Oven temperaturerC)
Fig. 2.8. Storage modulus (E') of UF resin after exposure to 120°C for various heating times with hot-platen heating.
In case of heating at 160°C, similar
E'behaviors were obtained. However, the resin pre-heated for 3 min was considered to have fully cured because the E' had an almost constant value during the entire measurement.
-(3) Dynamic mechanical properties of resin pre-heated with steam-injection heating
Fig. 2.9 shows theItbehavior of the resin after pre-heating at 1200
efor various heating times with steam-injection heating. Compared to Fig. 2.8, each curing process of the resin with pre-heating times was very different. In Fig. 2.9, irrespective of pre-heating times, each It at 200
ehad almost the same value, and the It curves during measurements were similar to each other, indicating the temporary decrease inIt values up to lOOoe. Judging from these behavior ofIt, each steam-injection heated resin was considered to cure to some degree regardless of the heating time. However, it was thought that UF resin did not cure
~=ngumes(m;nl• Untreated o 1 • 5 b,. 10 250 200 150 100 50 10·2L-.~_-'-_~---'-~_.J..-~---'--~---' o Oven temperaturerC)
Fig. 2.9. Storage modulus (£) of UF resin after exposure to 120°C for various heating times with steam-injection heating.
Pre-heating times (min) • Untreated o 0.5 b,. 3 250 200 150 100 50 10·2L-~--'--~--L._~_l...-~--'--~--' o Oven temperaturerC)
Fig. 2.10. Storage modulus (E') ofUF resin after exposure to 160°C for various heating times with steam-injection heating.
completely with steam-Injection heating at 120°C even at longer heating times.
Fig. 2.10 shows the
E'behavior of resin after pre-heating at 160°C for various heating times with steam-injection heating. For the 30-s pre-heated sample, the value of
decreased to some extent and then increased with increasing oven temperature. This result revealed that resin pre-heated under this condition cured to some degree. Prior to DMA measurement, the UF resin sample under the above conditions was slightly rigid which was thought to have contributed to the above result. However, for the resin sample pre-heated for 3 min, the value of
E'at 20°C was lower than that of the sample pre-heated for 30 s. The value of
E'continued to increase without showing a temporary decrease until lOO°C with increasing oven temperature. In fact, the sample state before measurement was very soft and wet. Consequently, as steam-injection time increased, UF resin was considered to cure to some degree and then return to a liquid state.
The dynamic mechanical curing behavior of UF resin under steam-injection heating was investigated using the mechanical curing index as described in 1.3.2 (4). The results are shown in Fig. 2.11. For the sample pre-heated at 120°C, the index increased gradually for 5 min, followed by a slight decrease with increasing heating time. For the sample pre-heated at 160°C, the index rapidly increased for 30 s, and then decreased immediately until it reached a 0 value after about 5 min. According to the result of the section 2.2, UF resin rapidly cured to some degree in the early stage. Afterwards, however, the reaction of the resin apparently came to equilibrium at 120°C, and hydrolysis gradually took place at 160°C. This result was slightly different from that of the dynamic mechanical curing index in this study. This was the reason why the dynamic curing mechanical index was calculated from the change of the
E'of the sample which includes not only the resin but also water. Therefore, we came up with the following conclusion: in the heating temperature
20 16 120 ·C ~---o 12 8 4 100 >< CD "0 .5 80 CI l: i : :J U a; U Cas .l: u CD E u
·eas l: >-C 0
Steam-injection heating time (min)
Fig. 2.11. Dynamic mechanical curing indexes for steam-injection heated UF resin as functions of steam-injection times.
-at 120°C, UF resin gradually cured to some degree in the early stage. Afterwards, although the curing reaction apparently came to equilibrium, the resin was slightly diluted by the by-products from the condensation reaction of the resin and the condensed water from steam. In the heating temperature at 160°C, the resin rapidly cured to some degree in the early stage. Subsequently, the resin rapidly returned to a liquid state because of the occurrence of decomposition such as hydrolysis and dilution of the resin with condensed water. That is, from a mechanical viewpoint as well, it was suggested that the curing behavior of UF resin under steam-injection heating differs much from that under hot-platen heating. 2.3.3 Summary
The mechanical curing process ofUF resin treated under several conditions was studied using DMA. The following results were obtained:
1) The curing process of liquid UF resin with catalyst proceeded through a two-stage curing process with increasing oven temperature.
2) The cured states of resins pre-heated by hot-platen heating and steam-injection heating were observed. In addition, a dynamic mechanical curing index was calculated to reveal the dynamic mechanical curing behaviors under steam-injection heating. It was found that under steam-injection heating, UF resin immediately cured to some degree, and then, dilution and/or decomposition took place.
PART 3 Bonding Performances of Wood Adhesives under High-Pressure Steam 3.1 Introduction
In the previous parts, the curing behavior and mechanical properties of several wood adhesives under high-pressure steam using a special reaction cell were discussed. In this part, the effects of additives and steam-injection time on the bonding strengths of PF resin and UF resin adhesives were investigated. In addition, the suitable bonding conditions of these resins under high-pressure steam were developed.
3.2 Bonding strength of PF resin
3.2.1 Materials and methods (1) Materials
The same aqueous PF resol resin as described in the section 1.2.1 (1) was used in this experiment. Wheat flour and sodium carbonate (Na2C03) used as PF resin additives were supplied from Oshika Shinko Co. Ltd. The glue preparation and their properties are shown in Table 3.1. Rotary peeled lauan (Shorea spp.) veneer wi th a thickness of 1.3 mm was used. The air-dry density and the moisture content of veneers were 0.56 g/cm3and 8.9%, respectively.
(2) Manufacture of plywood and measurement of bonding strength
Table 3.1. Preparations and properties of PF glue.
PF resin (phr) 100 100 100 100
Wheat flour 15 15
Sodium carbonate 2 2
Viscosity at 25°C (Pa's) 0.1 3.5 0.38 4.65
Gelation time at 135°C (min) 13.27 11. 75 8.02 11. 75
Fig. 3.1. Production method of grooved veneer. Note: The veneer was pressed for 40 s using hot-platen at 160°C.
the veneers. Therefore, grooves were set by hot-pressing in the face veneers as shown in Fig.3.1. The amount of the glue applied to the core veneer was 180 g/m2. 3-ply plywood
with the dimensions of 30 X 80 X 3.5 mm were prepared using a specially designed pressing apparatus as shown in Fig. 3.2. The specimens were subjected to press pressure of 0.3 MPa exerted by a pair of springs. The apparatus containing the specimen was then put into the reaction ce1l33) for steam-injection heating. Steam-injection heating was performed
according to the section 1.2.1 (2). The heating temperature was 160°C (steam pressure of 0.63 MPa) with heating time of 1-10 min. After the apparatus containing the specimen
---Fig. 3.2: Production of 3-p1y plywood using a specially designed pressure apparatus.
-was heated for the scheduled duration, it -was taken out and cooled by dipping in liquid nitrogen. As a control plywood, 3-ply plywood was also prepared using untreated veneers. The apparatus containing the specimen were heated in the oven at 160°C. In order to cure the resin completely, a heating time of 30 min was applied. The specimens were then cooled using similar method as steam-injection heated samples.
Intermittent steam-injection heating, that is ; steam is exhausted temporarily before the saturated high pressure steam fill in the reaction cell, then, steam is injected again for short period, was investigated. In this study, steam-injection heating for 1 min; steam exhausting from the reaction cell for 30 s; and steam-injection heating for 1 min, was applied.
The prepared 3-ply plywoods were conditioned at 20°C and 65% relative humidity for a week. Test specimens were prepared according to lIS Oapan Industrial Standard) K6802. After a cyclic boiling treatment (soaking in hoiling water for 4 h; drying at 60°C for 20 h ; soaking in boiling water for 4 h; cooling and testing in the wet state), the tensile shear bonding strength was measured. The total of six test specimens were used for each fabricating condition. The average value and standard deviation were calculated for each condition.
(3) Measurement oj glue line temperature
During steam-injection heating, the temperature in the glue line of 3-ply plywood was measured by inserting the copper-constantan thermocouples (wire diameter of 0.1 mm) into the center of the core veneer. The thermocouples were connected to a data processor via a data-logger, and the temperatures were recorded at a time interval of 5 Hz.
3.2.2 Results and discussion
(1) Effects oj additives and steam-injection time on bonding strength
The tensile shear bonding strengths of 3-ply plywood after cyclic boiling were tested by varying the additives and steam-injection time. The results are shown in Figs. 3.3-3.6. The percentages of wood failure in these 3-ply plywoods were, in general, low. This could be due to the considerably low pressing pressure exerted by the specially designed pressure apparatus. In the case of3-ply plywood glued with a neat glue (only PF resin), the bonding strength decreased markedly with increasing steam-injection time (Fig. 3.3). The maximum average value of the bonding strength was achieved in a steam-injection time of 1 min, which was almost the same as that of the control plywood. The bonding strength of the glue extended with a wheat flour also decreased markedly with increasing steam-injection time (Fig. 3.4). In addition, these bonding strengths were very weak compared to that of the control plywood. In the case of 3-ply plywood glued with the sodium carbonate added glue (Fig. 3.5), the average value of the bonding strength at 1 min was comparable to that of the 3-ply plywood glued with a neat glue at 1 min (Fig. 3.3). However, the bonding strength was increased with increasing steam-injection time, and the average values of